R help - Oct 2004 - Evaluate a function for various value of parameters

Hello list,

I have a problem ... and do not know how to solve it.
I would like create a function that estimate the quality of the fit of 
different functions with different values of parameters.
This problem is related to the following one:
I would like to create a function "evaluatemyfunction" which evaluate
a
function f for different values of parameters. For instance, if
f=function(a,b) a*b
evaluatemyfunction(f,1:10,5) would return:
5
10
15...

I would like that f could have a variable number of arguments. I have begin 
to write a function:

evaluatemyfunction=function(f,...){
b=match.call(expand.dot=F)$...
myg=expand.grid(lapply(b,eval)) #contains all combinations of parameters
for (i in 1:nrow(myg)){

argsasname=paste(paste(names(myg),myg[1,],sep="="),collapse=",")
# ??????
}

}

I have try to construct arguments of the function as character, then parse 
and pass it to f but parse doesn't work:

 >
parse(text=paste(paste(names(myg),myg[i,],sep="="),collapse=","))
Error in parse(file, n, text, prompt) : parse error

Any ideas ??

Thanks in advance,

sincerely
St??phane DRAY
--------------------------------------------------------------------------------------------------

D??partement des Sciences Biologiques
Universit?? de Montr??al, C.P. 6128, succursale centre-ville
Montr??al, Qu??bec H3C 3J7, Canada

Tel : (514) 343-6111 poste 1233         Fax : (514) 343-2293
E-mail : stephane.dray at umontreal.ca
--------------------------------------------------------------------------------------------------

Web                                          http://www.steph280.freesurf.fr/

Hi Stephane,

* Stephane DRAY <dray at biomserv.univ-lyon1.fr> [041023
00:20]:
> Hello list,
> 
> I have a problem ... and do not know how to solve it.
> I would like create a function that estimate the quality of the fit of 
> different functions with different values of parameters.
> This problem is related to the following one:
> I would like to create a function "evaluatemyfunction" which
evaluate a
> function f for different values of parameters. For instance, if
> f=function(a,b) a*b
> evaluatemyfunction(f,1:10,5) would return:
> 5
> 10
> 15...
> 
> I would like that f could have a variable number of arguments. I have begin
> to write a function:
That function has already been written for you. Lookup 'mapply' in the
online help.

HTH
Olaf

Stephane DRAY <dray <at> biomserv.univ-lyon1.fr> writes:
: I would like to create a function "evaluatemyfunction" which
evaluate a
: function f for different values of parameters. For instance, if
: f=function(a,b) a*b
: evaluatemyfunction(f,1:10,5) would return:
: 5
: 10
: 15...
: 

This is not a general solution (Olaf has already provided that) but 
in some cases you may simply be able to write your function f so that 
it takes vector arguments.  For example, with the f you have above 
one could write f(1:10,5) .

ÔÚ 2004Äê10ÔÂ23ÈÕ ÐÇÆÚÁù 06:10£¬Stephane DRAY дµÀ£º
> Hello list,
>
> I have a problem ... and do not know how to solve it.
> I would like create a function that estimate the quality of the fit of
> different functions with different values of parameters.
> This problem is related to the following one:
> I would like to create a function "evaluatemyfunction" which
evaluate a
> function f for different values of parameters. For instance, if
> f=function(a,b) a*b
> evaluatemyfunction(f,1:10,5) would return:
> 5
> 10
> 15...
> mapply("*",1:10,5)
 [1]  5 10 15 20 25 30 35 40 45 50
> outer(1:10,5,"*")
      [,1]
 [1,]    5
 [2,]   10
 [3,]   15
 [4,]   20
 [5,]   25
 [6,]   30
 [7,]   35
 [8,]   40
 [9,]   45
[10,]   50
is these meet your need?
try
>?outer
>?mapply
to see more details.

Gabor Grothendieck <ggrothendieck <at> myway.com> writes:

: 
: Stephane DRAY <dray <at> biomserv.univ-lyon1.fr> writes:
: : I would like to create a function "evaluatemyfunction" which
evaluate a
: : function f for different values of parameters. For instance, if
: : f=function(a,b) a*b
: : evaluatemyfunction(f,1:10,5) would return:
: : 5
: : 10
: : 15...
: : 
: 
: This is not a general solution (Olaf has already provided that) but 
: in some cases you may simply be able to write your function f so that 
: it takes vector arguments.  For example, with the f you have above 
: one could write f(1:10,5) .


Sorry, I missed the last part of your question where you
provided some code.  That code shows that you want all
combos, not just parallel evaluation of the args so here
it is again.

This answer creates a grid and evaluates.  It makes use of 
do.call and mapply. 

The first two lines of the body collect the args into a list and
create the grid.  The third line reestablishes the names of the
columns since expand.grid may not give the ones we want.
Lastly we preface the grid with the function name and mapply
the result.

	evaluatemyfunction <- function(f, ...) {
	   args <- list(...)
	   grid <- do.call("expand.grid", args)
	   names(grid) <- names(args)
	   do.call("mapply", append("f", grid))
	}